1 |
> The freeze module will covert the python code to c and then compile it. |
2 |
|
3 |
Sounds good in theory, does it work better that perl->c? |
4 |
|
5 |
> I think we can also look at using stripped down version of Python, like |
6 |
> stackless or pippy. |
7 |
|
8 |
Yes that must be evaluated. |
9 |
|
10 |
I guess I'm just very pro-python. We've developed |
11 |
> a great system that through the use of a very-high level scripting |
12 |
> language provides an easy framework for hacking and extensibility. I'd |
13 |
> hate to see that lost. |
14 |
|
15 |
As long as daniel lives, you whoun't see that lost. The plan is to |
16 |
convert parts of |
17 |
the ebuild python code to c++ once it is stable. Do you plan to hack |
18 |
things like the |
19 |
dependencie system? So finaly the core of ebuild will be c++ code used |
20 |
by a python module. |
21 |
Then it is possible to write modules for ebuild in other languages too. |
22 |
> |
23 |
> Since time is limited, why don't we just concentrate on what we've |
24 |
> always been doing? I figure that if can publish an API, then others can |
25 |
> develop ebuild in other languages. |
26 |
> |
27 |
I do not see a sense in using other languages than bash for ebuild's |
28 |
functions, but |
29 |
it whould be great to have a perl module from which I can access ebuilds |
30 |
functions. |
31 |
Of cause we currently concentrate on what we've always been doing. |
32 |
I'm reworking all our packages since two month to make intensive use of |
33 |
the USE variables. |
34 |
Once I'm finished we should be able to use USE as a global config file |
35 |
for all packages. |
36 |
So for example if you set "mysql" in USE all packages that can be linked |
37 |
against mysql |
38 |
get linked against it. I think this is usefull for embedded systems too, |
39 |
because it can |
40 |
reduce the package's size immense. There is no need for native languages |
41 |
support, pam or |
42 |
readline on an internet radio. |
43 |
Unfortunately all this is much more time-consumning as I thought and |
44 |
that is a main cause |
45 |
why there was no new release in the past. But once it is finished I |
46 |
swear that it will be real |
47 |
fun. |
48 |
|
49 |
> : BTW: Jerry, are you still headbanging or did you get gentoo working now? |
50 |
> |
51 |
> No. Unfortunately, I've become the latest dot-com casualty. |
52 |
Urgh, sorry to hear that, hope you get a new job soon. |
53 |
|
54 |
So, the |
55 |
> Vaio z505 had to go back. I went to a computer show over the weekend |
56 |
> and picked up the parts to put together a 1GHz Athlon system (I even |
57 |
> picked up a pcmcia bay so I can still try and get laptop portages made). |
58 |
> So, it's back to the drawing board for figuring out how to bootstrap |
59 |
> this puppy. |
60 |
|
61 |
I made a package for yard, which is a nice system for building rescue |
62 |
and install discs. Simplified |
63 |
you pass a list of binraies you want to have on a disc and yard finds te |
64 |
required libs and builds |
65 |
a working image for you. Debian uses a nice script to reduce the size of |
66 |
the shared libraries for such |
67 |
a disc by building the shared libs from _pic.a archives with only the |
68 |
required funtions. This can reduce |
69 |
for exanple glibc's size to nearly the half. |
70 |
The problem is that yard currently has the path to the config stuff |
71 |
hardcoded, so it must be hacked to |
72 |
read an env var or an parameter (should not be difficult). |
73 |
Once this is done, I think it's even good for making images for embedded |
74 |
systems. |
75 |
Additionaly I try to use syslinux for the boot disc. Hopefully this |
76 |
works better than our current system. |
77 |
Once more it's more time-consuming than expected but real fun in the |
78 |
end. |
79 |
|
80 |
All the trials we made for a week to get gentoo running on your vaio |
81 |
where tested here on an Athlon system, |
82 |
so hopefully it will work on your system too. |
83 |
Maybe I can upload new snapshot releases for build and sys this week. |
84 |
|
85 |
bye achim~ |
86 |
|
87 |
> |
88 |
> --Jerry |
89 |
> |
90 |
> name: Jerry Alexandratos || Open-Source software isn't a |
91 |
> phone: 703.599.6023 || matter of life or death... |
92 |
> email: jerry@×××××××.org || ...It's much more important |
93 |
> || than that! |
94 |
> |
95 |
> _______________________________________________ |
96 |
> gentoo-dev mailing list |
97 |
> gentoo-dev@g.o |
98 |
> http://www.gentoo.org/mailman/listinfo/gentoo-dev |